This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: TLS_INIT_TP, secondcall vs. USE___THREAD (was: GNU C Library master sources branch, master, updated. glibc-2.14-297-g52d4fef)
- From: Roland McGrath <roland at hack dot frob dot com>
- To: Andreas Schwab <schwab at redhat dot com>
- Cc: Thomas Schwinge <thomas at schwinge dot name>, drepper at gmail dot com, libc-alpha at sourceware dot org
- Date: Mon, 10 Oct 2011 09:42:18 -0700 (PDT)
- Subject: Re: TLS_INIT_TP, secondcall vs. USE___THREAD (was: GNU C Library master sources branch, master, updated. glibc-2.14-297-g52d4fef)
- References: <20110913083730.15283.qmail@sourceware.org><871uumj8z7.fsf@kepler.schwinge.homeip.net><m3wrcdjm3e.fsf@hase.home>
> > In the olden times, when USE___THREAD was still being defined (before
> > commit d063d164335938d557460bebaa7cfe388157b627), in both the two cases
> > above, TLS_INIT_TP(tcbp, 1) has been invoked.
>
> USE___THREAD was always 0 in rtld.
Not "always". But that was the status quo before the deconditionalization.
We still had some conditionalization leaving the possibility of __thread
or no for rtld itself, though nobody has done that in quite some time since
it's just simpler and more efficient for rtld itself not to use __thread.